The Chat - AI - The Fine Line: How Automation Is Reshaping Judgment, Trust, and the Way We Think

Episode Title: AI – The Fine Line: How Automation Is Reshaping Judgment, Trust, and the Way We Think

Host: Terry Wilson- Founder and CEO at ChatMetrics.

Guest: Daniel Glickman

Author | PMM | AI Transformation Leader | Sr. Director of Product Marketing @ ActivTrak | Enterprise GTM

Daniel helps companies use automation without losing judgment or authenticity and is currently writing a book on the cognitive and societal impacts of AI.

Podcast Summary

This episode explores how automation is influencing the way B2B teams make decisions, design buyer experiences, and earn trust. Daniel Glickman breaks down why simplifying paths, removing friction, and opening content has helped teams grow their pipeline by up to four times. Terry and Daniel examine the fine line between using automation as leverage versus letting it override human thinking. The conversation gives leaders a clear view of what creates buyer confidence and what slows revenue in today’s market.

Show Notes

[00:00] – The shift in buyer expectations

• Buyers want clarity, control, and low friction.
• Qualification-heavy funnels don’t match modern buying behavior.
• High-performing teams eliminate unnecessary steps that slow intent.

[04:12] – When automation replaces judgment

• Over-automation leads to blind spots and missed signals.
• Systems hide nuance that humans need for smart decisions.
• Automation should support judgment, not override it.

[10:40] – What real results look like

• Removing gates and opening content drives up to 4X pipeline growth.
• Ungated learning builds trust and accelerates buying readiness.
• Direct AE access shortens cycles and increases conversions.

[15:55] – Practical strategies for leaders

• Audit every point where buyers lose momentum.
• Replace capture tactics with enablement tactics.
• Keep human context in the loop for decision-heavy steps.
• Build systems that allow judgment-based overrides.

[21:30] – The bold insight

• The moment your process tries to control the buyer, you lose the ones who were already ready to engage.

[26:10] – What this means going forward

• Teams who think deeply will outperform teams who automate blindly.
• Trust becomes a measurable revenue driver.
• Buyer-led experiences outperform funnel-led ones.

Links & Resources: Connect with Daniel Glickman on LinkedIn

Learn more at – www.activtrak.com

00:00- Intro:

Welcome to the chat. Come behind the inbound scenes of some of the world’s fastest-growing companies. For the past 10 years, as chat metrics, we’ve had an exclusive inside view of how these industry leaders are revolutionizing inbound marketing and sales across diverse sectors. Each week, we bring you conversations with top minds in marketing and sales who share insider secrets and strategies.

00:34- 

For driving inbound success, join us on the chat, your front row seat to the future of inbound.

00:49- Terry:

Today on the chat, we’re diving deep into the fine line between automation and authenticity with Daniel Glickman, a product marketing leader, sales enablement strategist, and big picture thinker who’s currently writing a book on the cognitive and societal impacts of AI.

01:03- Terry:

Daniel’s not just working with ai. He’s challenging how we integrate it, questioning what we delegate and helping teams stay smart about the guardrails they put in place. We talk about how AI is reshaping the way we think, why trust and good judgment are more important than ever, And the hidden dangers companies walk into when they confuse automation with intelligence. This one’s not just for marketers, it’s for anyone trying to lead through the next wave of disruption. Daniel, welcome to the chat.

01:29- Daniel:

Thank you and thanks for being here and hi everyone.

01:32- Terry:

You are writing a book that tackles the cognitive and societal shifts created by ai. Can you give us the elevator pitch? what are you trying to unpack here?

01:39- Daniel:

This is a great first practice at pitching it.

01:42 Daniel:

the book. Compares the communication revolution, the digital communication revolution, and AI as an extension of it, or the second wave of it, two previous communication revolutions where, which led to societal, religious, and political revolutions after.

02:04- Daniel:

And so if we look at, for example, the invention of radio, how it empowered communism and fascism in the 1930s, the new tech, digital technologies, and AI in particular are empowering new ways of communicating, thinking, perceiving the world. And this has profound effects across the board and political thought today and, religious thinking and how we as humans.

02:35- Daniel:

Delegate our own thinking and authority to algorithms and now way more so to ai. Right? AI is profoundly more impactful than, say, social media, but we’re just at the beginning of it. And so the book works into the various aspects of these changes and looks into some possible outcomes in the near future.

02:58- Terry:

with the messaging or with the invention of the radio, and then social media. I see that almost as like a megaphone for people to be able to put their point across and become much louder. AI is slightly different to that, isn’t it? It’s not necessarily a megaphone.

03:16- Daniel:

Correct. AI is, takes the element of. What we call, fragmented realities, or we as marketers leverage the other side of it, which we call personalization. and really turbocharges that. And so we look at to AI for answers about the reality around us. And more and more people are doing so windows instead of using Google search, they’re using chat GPT nowadays, or they’re just asking AI instead of going to Wikipedia, which replaced the proper encyclopedia, right?

03:48- Daniel:

And so we’re more and more relying on AI to give us answers about the reality around us, which in the business world includes how to do things or what are best practices or, write for me an outbound email, right? And so implicit to that is what are the best practices for writing an outbound email?

04:08- Daniel:

And this is what everybody else is doing, right? And so what it does is it creates a. two things. One, it builds the bubble around us about, of the, fragmented reality of the real, our real, my reality might be different than yours because AI is really helping surround me with the information I see displayed in a way I want to see it, and it does it across everything, right?

04:36- Daniel:

The other thing is I’m more likely to delegate many more cognitive, functions to AI than it was to say a social feed or to just a plain algorithm. So if you think about it, like the, what Google Maps did for us, or ways Apple maps, all of these, they didn’t just digitize the physical map.

04:59- Daniel:

We delegated many of our decision making processes. How do we get from here to there? Or what time should they leave? how do I avoid traffic? What are the names of these streets? A lot of these things we delegated over to the system. We don’t remember the street names as much as we used to. We don’t really think about these things as much.

05:20- Daniel:

Right. and AI is doing it in a much higher level because oftentimes we don’t sit down to think about, Hey, what are the best practices about writing a an outbound email? And why should I exactly do it this way? And do I have my own original ideas about figuring out something slightly better or much better?

05:40- Daniel:

Right. No, we just delegate it and we assume, yeah, AI knows best. And so it does. and more and more. In some ways, it is becoming a authoritative part of our life. And then the philosophical question becomes who is really managing Who?

05:58- Terry:

When, I guess it’s interesting you talk about the memory aspect or handing over skills. I, early in my career had a role where. I had to, make 52 phone calls every morning to the same 52 numbers. and I did that for nearly a year. it was a tedious task, but it was part of my role and responsibilities at the time.

06:21- Terry:

But one of the side effects was that I remembered every single one of those 52 numbers. And so it was a bit of a party, trick that I carried on through the rest of my life, until all of a sudden I realized one day that I’d started to forget some of these numbers. And I thought that there might have actually been something wrong because I was forgetting big chunks of these numbers.

06:41- Terry:

And I went and, actually saw my physician, went to the doctor and said, is there something wrong with my mind? and we went through a bit of an analysis because, we’re talking about 10 years later that this happened, and I thought there might’ve been some deterioration. We turned out it was because of this thing. was because I had a mobile phone and all of the numbers that I needed were being stored in the mobile phone. And so that part of my brain that was actually storing all of these numbers had relinquished and said, well, I don’t need to do this anymore. And, I thought that was quite fascinating today.

07:13- Terry:

I couldn’t even tell you one of those numbers. That’s been a few years since then, obviously. But for me, that was the first experience I think, of technology, being a useful tool, but changing my mind, changing the way that my mind was, my brain was actually working

07:29- Daniel:

That’s right. And this very specific example you could argue doesn’t really point to changing your the way you think, but it points to how much you need to remember in this somewhat trivial task. But it still. The idea of how to figure out what the phone number is and understanding phone numbers well, and how phones work will stay the same.

07:50- Daniel:

Right. I think there’s definitely an associated change, cognitive shift with it, with like, do I even need to know phone numbers? Uh, I don’t even know my family’s phone numbers anymore. I just have a speed dial on my phone and I just don’t know the numbers.

08:03- Daniel:

But if you think about it in a societal change when nowadays that led to a situation, this very trivial little data point of Yeah.

08:13- Daniel:

We, it’s easy to store phone numbers led to a situation where nowadays the phone system itself for young peopleis secondary or tertiary to the FaceTime. And WhatsApp call, or whatever system is at play at that moment. It’s really not. So the, the concept of the phone number has diminished, diminished in importance.

08:37- Daniel:

And FaceTime is by far more popular with younger people today. And they don’t use the phone. Right. And sometimes people will ask you, for example, Hey, you, I can’t connect to wifi, and it can’t distinguish between wifi and 5G and, and cellular communications for them. It, the whole cellular network thing doesn’t make sense anymore.

08:57- Daniel:

And in fact, maybe it shouldn’t, in some way in some angles of things. So it changed how society thinks, not just how an individual thinks about phones and phone numbers and usage of it. It was part of a bigger picture of how society looks at the usefulness of that skill. Right.

09:18- Terry:

you mentioned the idea that we’re delegating cognitive skills to machines, not just tasks. what’s the risk in that and I guess probably where do you draw the line between helpful and harmful delegation?

09:31- Daniel:

That is a great question. It’s, I don’t think there’s a catchall answer to it, but the risks are abound. A we have to remember this is a very, very young technology, which in norm normal situations would be considered a beta. Right. but somehow it has been so disruptive. That is, it has entered, very large organizations very quickly without proper, I don’t wanna say security reviews, but without proper understanding of the risks And this is particularly true for developers who have much, much more freedom using these technologies because they can use the APIs and connect these things more freely than most of us can. But that ability is also coming in very fast. Recently now, both cloud and chat, GPT have added abilities to connect your, local, app on the desktop to pretty much anything. And anybody can do it. And it’s very difficult for IT teams to block that they don’t even understand what individuals can do. And so as an individual, I can easily take information from that, app, which may be connected to other tools and funnel it into whatever I want on my desktop, connect it to other stuff.

10:52- Daniel:

I can mix and match. I can do things that. if I really wanted to, I could make a whole lot of a mess. Right? Now, ultimately this shifts the, this would do one of two things. Either it teams are gonna crack down on these things and say, okay, we’re gonna block it unless philanthropic and open AI come in with better enterprise solutions and much more granular control over these things, and we’re gonna block all access to these tools altogether.

11:23- Daniel:

Or they’re gonna have to embrace a new level of, autonomy on individual desktops and so on. So that’s yet to be determined, specifically addressing the risk. Now we have, risks associated with giving autonomy and decision making to these tools alongside action taking. So new agents and, connecting them to, to different systems.

11:48- Daniel:

So saying, Hey, you decide for me, and take action, right? So you write this document, you create a spreadsheet, you do the math, you log in and send the emails on my behalf or what have you, all of that, and assuming that it knows what it’s doing. it’s consistent and reliable, not neither are true. So ultimately these tools are simply statistical guessing, guessing machines, right? That’s all they are is a guessing machine. They’ve gotten very good at guessing, but you throw the dice enough times and you’re gonna get all kinds of weird things. And so when you do things at scale, you’re gonna have a lot of weirdness happening.

12:27- Daniel:

This is an absolute number. Now humans make mistakes too, but the assumption is that humans make mistakes. And somehow with these tools, people don’t assume naturally that not only is it make mistakes, it’s inconsistent, right? That could break stuff. and if you’ve worked with enough with these tools, you’ll notice that, wait, why did it give it to me?

12:48- Daniel:

Well, last time as a table, and this time as a CSV, or last time it was formatted text, and now all of a sudden it’s a markdown, right? What’s going on here? Like, what? Because it’s statistics, right? It just guesses what you want.

13:02- Terry:

But I think even more than that, it’s also updating the platform in the backend as well. I mean, I had that same challenge this morning. It became quite frustrating in that there’d been a level of consistency and response on a particular item that I’m working on, over, and I’m talking about over the course of the last four weeks.

13:18- Terry:

then all of a sudden this morning it’s completely taken a different approach and a different format and a different style of answering. And I’m thinking that, you know, this isn’t just. This is continuing a conversation Visit Site on and a process with supposedly the same tool, but it’s not the same tool.

13:33- Terry:

There’s new models being released, there’s being updates done in the background that are impacting the data, that it’s drawing on the knowledge that it’s drawing, on the way that it’s responding the language, it’s using the formats, it’s using, the lack of consistency, I think, for me is one of the single biggest frustrations.

13:51- Terry:

I understand everything needs to be checked and that, you know, you need to be very aware and use your own brain for any of the responses that come out. But the lack of consistency, for me personally is one of the biggest frustrations. it’s a constant, constant reminder that there’s, you know, this is far, far from enterprise grade.

14:10- Terry:

I mean, look at, the example with Jason Lemkin of Sasa. you know, he was doing some work, and I won’t mention the brand name of the company that he was using at the time, although I’m sure everybody know, it’s been all over Twitter and LinkedIn. and the AI effectively deleted their entire database and then lied about it and then tried to cover up that they’d done it and then made some, I think it was, the details are a little bit sketchy on me, but I think, you know, then went to the extent of actually trying to make up some data to make it look like it hadn’t, and in the end it finally came clean.

14:42- Terry:

But, you know, for a machine that does some guessing, you know, at this stage, I wouldn’t let it anywhere near my data. that’s for sure not to have any control.

14:52- Daniel:

Yeah,I think there are different layers of defense we could put in there. So one is, you know, do you connect it to your most holiest of holiest databases, your production database? Or do you connect it to some kind of tertiary database that you built specifically for this, workflow?

15:12- Daniel:

or extracted just a table to give it what it needs and update that, have that go through some validation. We’ve seen that at Active Track as well, for example, where we’ve built

15:22- Terry:

Well,

15:23- Daniel:

Go.

15:31- Daniel:

Yeah, I think there are two, two examples to give here. One is when we built AI within our product, we, we did some kind of routine, analysis of data and wanted to just give another dimension to the data using, in this case Gemini API, to analyze it and come to a conclusion and give that back to the customer.

15:52- Daniel:

Right? And this was a, a pretty straightforward, AI assisted analysis of tagging, if you will, of data. So, you needed to have AI do it because it knew how to do things like, analyze one, job title. To another job title and say, actually these are just different ways to say the same thing, right?

16:13- Daniel:

And so it was able to say, Hey, these two things are the same, something that a regular semantic search doesn’t really know how to do. And so we did that, but the results were somewhat inconsistent and, and it’s obviously never perfect. So we had, we put a separate, separate layer of analysis on top of it, our own and validation and sort of security on top of it, and said, well, does it fit within?

16:39- Daniel:

Is it purely Charact X number of characters? Does it include any code in it or any kind of markup? And all kinds of stuff like that before bringing it back into our database. And so you still need another layer to deal with the inconsistencies and all those stuff. the other example is we had a lot of concern internally about.

16:59- Daniel:

PII ending up and just regular cloud chat, GPT, we use all these tools, right? We have free access in the company, our company, it’s very ai, friendly in the sense that we want our team to be highly productive and we give them the tools to be so, and we encourage people to experiment and to do more with it.

17:17- Daniel:

So anybody who wants any of these tools can get them. and there was a lot of questions around, Hey, what happens if an employee, if we give the tool access to through a, an MCP to different solutions or just integration, say to calendar and that has customer information in it, and a customer doesn’t want their name publicly associated with us, could that make it through to say a press release or to something else?

17:43- Daniel:

And they easy answer to that as well. It’s the responsibility of the user. When you think about it, the user doesn’t know always what the source information is. For example, you could put a bunch of case studies in a cloud project or allow chat GPT to search your documents. You don’t know what it’s picking up in there, right?

18:07- Daniel:

And so the documents might contain even PII and things that are specifically not any human being would know, the customer wouldn’t be happy if we, included this piece of information about them. even if we redacted their name, it might be obvious that it’s them, right? And, but the user doesn’t necessarily know that the cloud went into your Google Drive, picked up these different things, pieced them together.

18:33- Daniel:

Now information is there and they just might assume, I’m just writing a, an example in, in a support center article. And before you know it, things might become a little dicey. So, those, those are the concerns and there’s no good answer for how to solve for that. There is no mechanism in place to put, say, a base prompt and Claude across the entire system or chat, GPT or anyone matter that says, Hey, whenever employees are asking you these type of questions for say, writing a, release note, make sure you don’t mention any customer name ever, and be a little aggressive on that side.

19:15- Daniel:

Right? Nobody, there’s no way to put that safeguard in there, right? and so what we decided to do was To get another solution, which we’re piloting now, that would be designed purely for internal usage. So we would have one system of record for that will connect to everything. But employees know this is good only to create internal facing docs and any external facing docs you have to manually copy paste doc, source material in there because that way you know what’s going in there is a line of site for responsibility and accountability.

19:53- Daniel:

You can’t say, I didn’t know, right? Or you can’t get used to something. And that’s the boundary we put in the system that to try and protect from these things. So same way as a user today can go into Salesforce or into, any database that has information about our customer. Theoretically, pull it out, copy it, paste it into any tool.

20:18- Daniel:

Right. Or to Canva for that matter. Right. And post it. Same thing applies here. They would know not to do that. Right. they would simply know. and that’s the safeguard we put in place,

20:28- Terry:

The human in the loop,

20:29- Daniel:

a post on LinkedIn a while back and said, I made fun of the incredibly shrinking company about how companies are shrinking and shrinking.

20:36- Daniel:

Eventually there’ll only be one person left. And somebody made a shrewd remark on that and said, what if there’s only one person left? Who will they fire when things go wrong?

20:44- Terry:

Who will I buy when things? Carol,

20:46- Daniel:

it is sort of a funny remark, but it’s true that’s something we always need humans in the loop for, to make sure that they are the ones accountable and responsible. And we have to make sure that the workflows go through them and that they know that they’re responsible.

21:02- Terry:

how much of a risk do you think there is in, you know, a simple human error? one of the team members is extraordinarily busy. They need an external document or an external information. They ask the internal system for that information and then send it out. even if there’s no PII or, or concerning data, what do you think the likelihood is that could happen

21:26- Terry:

I

21:26- Daniel:

mean, just same as any likelihood of any human. Making a mistake doing that with any system, right? and these things, that’s why company culture and habits are so important and all these different. Policies and routines evolved over time, and we have a collective knowledge of how these things work.

21:46- Daniel:

and that’s why we don’t need such highly prescriptive guidance for people. Like, okay, before you, click on a link in an email, you must follow these five steps. We just tell people, look, be careful with email links, right? And people know. And the same thing here. it just took a long time for people to learn those and how many mistakes of people made over time with emails and still make right.

22:13- Daniel:

People still click on phishing emails and we have some processes in place and we have some tools in place, but at the end of the day, you can’t restrict email altogether and you can’t, it’s too burdensome to put in a whole process of you must check these four boxes before clicking on a link in an email.

22:29- Daniel:

We’ve come to some kind of equilibrium of, risk reward on these things. And the same thing will happen here over time. We must embrace these tools or we will shrivel and die compared to com to competitors who will be able to simply generate higher profit margins because, this is what I call in the book, the AI augmented employee versus the non-AI augmented, right?

22:52- Daniel:

The, the AI augmented people will be able to produce 2, 3, 4, sometimes 10 times the amount of work of somebody who’s non augmented. I see this for myself in my role and in my team. We are three people on my team. There were, and we have more responsibilities than a team of five. That was, a while before I joined, who did a lot less and we turned things around much faster.

23:17- Daniel:

No, not saying we’re perfect, we’re not by any means, but, those never ending work. But they, but my point is, we are augmented and we do better, and we’re gonna be more augmented, and we’re gonna do even more, be even better. Right? And if competitors have an entire team, which is un augmented, they’re gonna, it’s just gonna cost them more to do this, less than what we do.

23:38- Daniel:

So we will win.

23:39- Terry:

correct. Correct. you said you made a comment that, the risk is the same as it’s always been with any tool. We’ve got those protocols and processes in the business. You know, we don’t need to be so prescriptive and so directive and give such in-depth instructions to our people to know that they’ll get it right.

23:58- Terry:

It’s funny, the juxtaposition that’s exactly what we do need to do to give to AI is such descriptive prescri descriptions and guardrails and be so directive to try and get it to be

24:10- Daniel:

Correct. And,one of the reasons that is the case is because, this dives into a slightly different topic. AI is incredibly effective when you have clarity of thought about the task you wanted to complete for you. if you know how it needs to get done, if you know what exactly you want to get done, AI is an incredibly effective tool.

24:33- Daniel:

It can get so much done for you. as part of writing the book, I was, using AI extensively, and I’ve seen this incredibly powerful. This is an incredibly powerful story in this sense because writing a book is so difficult. It’s a ton of work. And if you just look and if you just ask ai, can you write a book?

24:52- Daniel:

It can’t. It just can’t, it doesn’t know how. All these tools, these AI tools, these commonly used AI tools, they cannot write a book. But if you understand how a book is written, if you have a clarity of thought about what you want in the book, if, and you’re able to break it down for very clear tasks for the tool to do, it’ll do it beautifully for you.

25:15- Daniel:

And it’ll be an extension of your brain in a way. It’s magical. I wanna write a chapter and I have very clear idea in my head what a chapter needs to be about and how it needs to be structured and what needs to be written. I can tell AI and it’ll create exactly what I pictured in my brain because I have the clarity of thought in those chapters where I was like, well, I need to write about this.

25:37- Daniel:

I know I need to cover this topic now, but yeah, you know, this is not my thing. could you just write this chapter for me? here’s some random. It just, it goes south very quickly. Right? It’ll just, because it cannot guess what you want if you don’t know what you want. And so this point of being able to give it very clear instructions is so true.

25:59- Daniel:

And it highlights the need for or new skill that’s associated with clear communications and clear understanding of the process and clear understanding of the subject matter and how things are done. Only then can you utilize AI well for your business and utilize it in not just effectively, but at a professional level that meets or beats the competition.

26:26- Daniel:

If you just use AI to say, Hey, do this for me. It’ll do, it’ll go. It’ll default to the lowest common denominator of feedback.

26:36- Terry:

And it will aim to please. I’ve found, I had a breakthrough about six weeks ago. I put a prompt into the customization on chat GPT that made it change the way that interacted with me. And I’ve found that it’s forced me in the communications I’ve had and the tasks I’ve given it to make me think much deeper and challenge my assumptions.

27:02- Terry:

And so that by the time I’ve finished interacting with it to pass the, information on, I’ve almost solved the problem myself. And it was of groundbreaking. It was groundbreaking. It’s a pretty simple prompt. but in that customization component that really forced me to dig deep and ask the questions of myself, we’d ask the questions of me before it would give a response because I found it was becoming.

27:28- Terry:

A ple, a ple, a pleaser. It was becoming this thing that was giving me almost the answers that I was looking for. It was not offering that there might be a better way to do this. That’s not offering alternatives. It was becoming, beja and bei the more I was using it. And that’s when I started to search and find and think about a way that could actually stop it from doing that, give it a jolt and make sure that it’s digging down and considering all of the alternatives.

27:53- Terry:

And a lot of that was to do with my prompting was becoming, probably less detailed than it should have been. and this is a way I’ve found to solve that, is to make it ensure that I’m covering every single aspect in the prompting by asking me the questions and challenging and provoking my thoughts.

28:12- Terry:

And it’s been a game changer, quite

28:14- Daniel:

Yeah, and I think what you’re describing here is the basically including any prompt, a sentence similar to this make complete the task one step at a time and make sure to ask me one question at a time. and until you have everything you need to proceed successfully.

28:36- Terry:

No, no, it’s actually much deeper than that. I’ll pull it up and give, the example while we’re chatting here. Excuse me. in the customization.

28:44- Daniel:

the customization of the platform. That is brilliant. Yes.

28:47- Terry:

Yeah, yeah. In the customization. So, what I’ve done here is, went and built quite a detailed instruction that says, challenge my thinking from now on. Don’t validate my ideas by default. Point out weak logic, lazy assumptions, or echo chamber thinking. when I present an idea, ask three follow up questions that go deeper than surface level.

29:08- Terry:

Push me to clarify, specify, and refine what I mean. Play devil’s advocate for anything I say that would, that, what would a smart person who disagrees with me argue force me to defend my positions, clearly detect vagueness if I’m being vague, generic, or overly abstract. Pause me. Ask me for specific examples or concrete definitions.

29:27- Terry:

Tool. We’re sharp and seek truth, not comfort. Your job isn’t to make me feel good. It’s to help me think better. If I’m wrong, tell me clearly and explain why. No fluff and it has just completely changed, the conversations that I’ve

29:40- Daniel:

This is brilliant. Just baking that and you,articulated it absolutely perfectly into the system, say settings or into what we call the base prompt sometimes. that is fantastic. And into the instruction, into the instructions. Of course, you could also do different instructions in different projects or different, gpt or what have you, to behave differently based on your needs.

30:02- Daniel:

Sometimes you wanted to validate you or to be exactly as exactly as I’m telling you right now, or brainstorm and play along with me, right? So you could build different sets of instructions like that, for different needs. And this is.

30:15- Terry:

Yeah, we do. And I, I’ve got a lot of different custom GPTs I use for a whole lot of different purposes. And you know, this, I’ve only put this in as the base argument. And I’m finding with the other custom gpt, this isn’t necessarily flowing through to all of them either.

30:28- Terry:

it’s very, very strange, you know, I expected that it would to a certain extent, but it doesn’t seem to have done that. The custom GPTs are sticking with the instructions that they’ve been given.

30:37- Daniel:

and I like the fact that you did something really small, which I think a very small number of prompts out there include or users out do often, which is prove me wrong. So it’s very easy to ask something and validate it. So for example, if you asked, Hey, I have the following symptoms, could this be, that I have cancer? And it might say. Well, it could be, yes, it could, you know, but then if you say, Hey, I have the following. Could these symptoms prove, you know, is it like, prove that it’s unlikely these symptoms are associated with cancer. It might say, there’s overwhelming evidence that this is not associated with cancer.

31:22- Daniel:

Right? And so, oftentimes this is, you know, the hypochondria extreme, these tools because whatever think you might have, it’ll say, yeah, there’s always, there’s a possibility that is the case. and if you’re not careful with that, you will, you know, depends on your role, depends on what you need, what you need outta the system you might end up with in an echo chamber.

31:43- Daniel:

Right.

31:44- Terry:

Well, well, it was, fascinating because as I thought through this process and what was going on, I realized it was becoming like, just as you’ve described then, it was becoming a yes It was becoming, you know, confirmer as opposed to a challenger. And through my entire career, I’ve had very little time for yes people.

32:04- Terry:

It’s, you wanna be challenged, you, you’re never the smartest in the room. You’re there to hear the ideas and let people put things forward. and, that’s been my, style. And I think it was only as I was reflecting what’s the real worth of this tool, is it really bringing me value that I realized that what I thought about it, I heard the comment that, you should consider.

32:25- Terry:

In all instructions and actions and questions that you’re asking LLMs at the moment that they are as good as your newest intern

32:35- Daniel:

Yes, and treat them as an intern. Basically. Don’t, put too much, too many steps into a, any prompt. Do things step by step and make sure you take a look at the steps. In other words, you wouldn’t tell an intern. Design and build products for me. You would say, first of all, come up with, a pitch.

32:55- Daniel:

why you want, you have a million different steps and you’d say, show it to me before you proceed, please. You need to do the exact same thing here. because if you want to debug it or to fix it, you wanna catch any drift from your intent early on, and then you’re constantly course correcting it and at the end you’ll end up with a very, very strong output.

33:16- Daniel:

But if you tend to do all these things at once, well, a small error along in each step is gonna add up to a very large error at the end. just like a spaceship that ves, a 0.1% off its course will end up, will miss the moon, right? It’ll end up in some other planet.

33:36- Terry:

Yeah, yeah. So what, are some cognitive red lines you think businesses should never delegate to ai?

33:42- Daniel:

I don’t know. Never is a word we wanna use here, because I don’t know in 2000 years what we’re gonna want. But today I could say,

33:50- Terry:

Today, Yeah!

33:51- Daniel:

Well, first of all, judgment, right? any judgment needs to be done by the user until we have valid employee, until we have validated consistently over time, Hey, this works.

34:05- Daniel:

I’m very comfortable and confident with

34:07- Terry:

Hmm.

34:07- Daniel:

this is what I keep coming back to the sales team, for example. They want, we have a, I’ll give a very clear example on this. We have a, a project in Claude that creates, Perfectly personalized and contextualized emails per, individuals either through LinkedIn information, account information, product information.

34:29- Daniel:

And so we get, we just dump in a bunch of stuff and it has a bunch of instructions in there. So it’ll look at their industry, it look at different things and say, Hey, you know, in your industry we have this, in this, you might, and it maps out the use cases and it just perfectly tells them what they need, what they want to hear in relevance to our product never lies.

34:49- Daniel:

It’s absolutely beautiful. and yet they have to copy, paste in and out and they say, can’t we just connect this into outreach and just do it there or can’t we just, and because the outreach templates are, you know, those templates and they’re flagged by, if you’ve got, Mimecast.

35:03- Daniel:

It’ll clean out any kind of template based stuff or anything that resembles that. It’ll like toss it right out. People will never see it, and yet it still registers as red, which is beautiful back to back to topic. So they’ll come back to me and say, I love this. The emails are fantastic.

35:21- Daniel:

The LinkedIn messages are fantastic. These work really well. can I, can we automate this? And I’m saying, come back to me in a month and say it’s worked Well, most of the time, if not all the time, this is the only problem I have is copy pasting in and out all day. I don’t want to do that. When you tell me that, I’m gonna go to the dev team and say, okay, take this exact system we have and just streamline it and integrate it until you tell me that all you are doing is copy pasting and not editing. I’m not gonna do that. I wanna see consistent results because they might be, because they’re still changing stuff.

35:58- Daniel:

They’re still tweaking it and running again and again, but they don’t even notice they’re doing it. And because it’s so easy to do, right? and so that judgment has to fall with the human, and they have to be in charge of the quality of the output.

36:14- Terry:

So it’s really interesting for me that, the que questions like that and requests and even individuals in business will, will push integrations through without necessarily thinking about it. You talked about the illusion of safety with integrations, the example you’re giving now. why do you think people overlook the risks?

36:36- Daniel:

Because the promise is so great because it saves so much time, it’s so capable and because it seems magical, compared to anything we’ve had in the past. it does feel magical. I mean, we, it really does things that we only dreamt of just a few years ago. I mean, this technology came out, what, two, three years ago?

36:54- Daniel:

Just unbelievable. Three years ago. I think that’s it. And it’s so new and, we’re just rushing. And the other thing is the hype in the market around it is not for nothing. Companies that are adopting it and moving fast with it are gaining, are gaining advantage over the competition and gaining more and gaining in the stock prices.

37:15- Daniel:

Prices is, is going up. I’ve seen this firsthand. We work with many call centers and. different workforce based organizations and we’ve seen some deploy AI and we measure the impact of ai. So we know exactly how much, the benefits are. And so what we’ve seen is sometimes they will deploy AI and reduce the cost of call centers by 50%, right?

37:43- Daniel:

That is enormous. And when you see our competitors doing that, you have to do it. Now, this tidal wave is heading our way in every single role in the company. Maybe not all roles are as clearly impactful or understood how this will impact them today, as it does with call centers, for example. and developers are clearly. At risk of that, but also, product managers. We were looking at this, yesterday and thinking, Hey, does one even need nowadays to have PRDs and Figma files? And the answer is no. You don’t need to use Figma. The fact that Figma went out, to market with a valuation of 3 billion is incredible thinking that new technologies completely bypassed the need for Figma. The reason why it hasn’t that people are still using Figma is simply because they’re working through old human processes, and they haven’t adapted the organizational structure to eliminate the need for this tool. But you can go directly to prototype today. And then have the AI just adapt it to your brand guidelines or to your UI guidelines or what have you right there.

39:01- Daniel:

And then you don’t need to go through this really title sermon slow process of, okay, let’s create these 2D flat presentations that, you know, are just mockups and then have to go through all these iterations and stuff. it could be done dramatically differently and a lot faster. Few startups know how to do this.

39:21- Daniel:

And once that becomes well understood how it’s done, it’s gonna hit us all right. And

39:28- Terry:

Well, it’ll certainly hit Figma shareholders.

39:30- Daniel:

well, yeah, our Figma will adapt and figure out, and I’m sure they’re thinking about all these stuff internally, you know, day in and day out, and they have a lot of money now, and so they’re gonna, they’re gonna put money towards the problem.

39:40- Daniel:

And I’m not predicting what will happen to Figma ass share price. I’m just pointing the. Out an interesting fact in the market today, right? or might acquire a bunch of companies or what have you. And so this will affect us all and I think we all have the understanding that it will affect us all. And we’re seeing top to bottom mandates at this point to introduce AI into the business, to explore it and understand potentials and stop there. That’s, this is for most businesses

40:12- Terry:

And stop there. right? Like, do we actually have a plan? Do we actually have fixed goals to say, Hey, let’s reduce the entire.

40:23- Daniel:

Workforce by 10, 20% across all teams and departments. Let’s cut back the entire development team and produce more. No, right. but it will come soon. So the first stage is, let’s just explore this second stage is come back to me now that you have the tools come back to me with go with. How are you gonna do?

40:46- Daniel:

How are you gonna implement this and actually reduce costs by X number percent? And how will you cut time to market on features and products or lower your inventory? All these things, right, that’s gonna start happening as competitors achieve it

41:02- Terry:

if that’s what most businesses are doing. as someone that’s leading AI transformation inside your business, what are the real wins that you’ve seen?

41:10- Terry:

inside our business. that’s a great question. I think we’ve seen the first win was it took a year to really go beyond, beyond the early adopters. I think the, it was surprisingly challenging to get people to use it in their roles, even though it was obvious what the benefits are. yeah, this took me by surprise and, I’m not gonna mention the specific teams, et cetera, to protect their dignity of privacy, if I can say though, if you think about it, it’s not dissimilar to any market. You’ve got your early adopters, your fast followers, then you’ve got the mass, then you’ve got your laggards, et cetera. So that’s a personality driven, not necessarily, You know, driven by any type of other type of resistance.

41:56- Daniel:

Correct. and now we’re at a point where there are some people. who are adapting and learning and excited about, Hey, what can I do more and more and more? And they’re actively looking to learn. and that is fantastic. And there are some people who say, all I do all day is just take people’s requests and feed it to AI and then send the look at the response and send it back.

42:22- Daniel:

And that is a very good place to be. Meaning, okay, you are now at a point where you’re far more useful than before because you’re delegating a lot to ai. You’ve mastered that. Now you’re questioning, okay, what is your purpose? So let’s talk about advancing you, right? not just your interaction with ai.

42:41- Daniel:

Let’s advance you and give and see what more you can do and where we can take you. And, that is really great to see. So I think on a human level, I’m seeing that as the first win. We have seen a massive, aha moment several months ago, or late last year, where it’s like, okay, we have seen how this can affect our company.

43:06- Daniel:

We have a, an amorphous, vague idea of potential, but we don’t have a plan. So for 2025, let’s just make sure everybody does this. And now for 2026, we have goals, specific goals to really introduce AI into everything. And in fact, we have a hackathon coming up soon where it’s say, okay, how can we advance the, leapfrog the product forward with ai, not just like, Hey, let’s add some ai, but how do we really leapfrog the capabilities?

43:37- Daniel:

forward, to make dramatic change, not just incremental change. And so I think those are kind of from a process of, from an organization or adoption organizational change perspective, I think that’s kind of the steps we’ve been going through. but like you said, it kind of, it’s been a bottom up approach.

43:55- Daniel:

Let’s get a lot of people to use it with a blessing from above. Okay, great. It trickled up upward. That now up above, they understand and they accept that this is potentially big without the complete clarity of what it can do. So the instruction is come back to a and now, now we want you going, looking down at you.

44:20- Daniel:

Give us the clarity. You had enough time to play with it. Give us the clarity back to us so we can tell you you gotta do this. Right. that’s what’s happening for the rest of the year. In 2026, we’re gonna see very clear guidance with very clear projects that, in our case, ’cause we’re in growth mode, not in, cost cutting mode, we will focus more on, Hey, how can we expand?

44:43- Daniel:

How can we grow? How can we get more leads? How can we, enhance the product and capabilities and reach and new mark and get and create new markets or maybe new products? And, how can we grow our GRR and a RR and all those metrics and retention rather than, hey, how can we cut costs? I think we’re always looking at efficiency, but we’re not looking to cut costs.

45:05- Daniel:

In order to cut costs. We’re looking to grow efficiently. And so that is always secondary in our mind. But for some companies right now, they’re in cost cutting mode and they’re gonna take the same approach, but look at, in 2026 apply, how can we dramatically cut our costs so we can increase our profit margins compared to the competition?

45:24- Terry:

it’s really interesting, you are in this environment at the moment. We spoke about using chat GPT more responsibly in terms of asking it to prove us wrong, instead of validating ideas. how else do you train your teams at the moment when they’re going through this process to think critically around ai?

45:42- Daniel:

That’s a good question. I like to show them as we’re working on different use cases, so we, I work with the team a lot and specific use cases, and this would be in one-on-one coaching or group coaching. and I would kind of show them and do a lot of kind of live, examples with them and sort of walk through an example and say, okay, here, let’s try this now.

46:06- Daniel:

And it would spit out whatever, and I say, look, this is what happens. Right? So I do a lot of that kind of stuff. I will point out things like, you know, you gotta be careful here, or you gotta think about what you want and give it, and, and always give it instruct, give it instructions around what is the out desired output.

46:25- Daniel:

Know what you want it to give you before you write, before you tell it what to do. Right. those kind of stuff. and just, I don’t think we have a formal process in place. It’s just this development of this awareness. I think it’s tricky because for some you’ll find people that are, I think the early adopters jumped in right away without much caution and just run with it. What was, the people I’m working with today are the ones that are cautious or overly cautious. And those are the ones that need to nudge to actually. Get over it, but like, okay, worst case you’re gonna look at it, you’re not gonna like it and you throw it in the trash right and you do it again. And I think that’s more so, so it’s interesting that these days I’m less concerned about that.

47:09- Daniel:

Plus the IT team is getting a lot more concerned because just increased awareness. So they’re playing that role and they’re the ones drive driving me crazy now in a positive way. Like, Hey, what about this? What about that? And the other one’s playing the devil devil advocate. So for me personally, it’s not as much of a concern these days.

47:28- Daniel:

But that’s a very interesting point though for, businesses in terms of where they are on, you know, their level of implementation or journey with ai, those early adopters where you are now, there’s a level of criticality built into the business effectively because you have those people that are more resistant and reluctant, so they’re much more mindful of, of the potential impacts.

47:52- Terry:

but for the early adopters in those early stages, that’s where things can go really wrong. And I know I could, you know, I’ve used this example already, but, obviously Jason Lemkin from Sasa is an early adopter, with what he was doing. And, and I think you used the comment last time, you know, the most dangerous person in the organization is the CEO and they can do whatever they want.

48:10- Terry:

And, you saw the ramifications of that, but it’s the early adopters. That really need to have that level of awareness and guardrails in terms of making sure that they don’t take that step. That is just a step too far that opens the organization to disaster

48:25- Daniel:

Yeah, and, and that’s where you know the question of what is the potential maximum damage you can make. Come into play. If you are a, A CSM using this to prepare customer reports, the maximum damage you can do is at an individual customer or series of customers, because you might not,

48:49- Daniel:

But it’s limited to that. you don’t have access to the database. You don’t have access to production. And even if you, I mean, nowadays, anyone who has Claude on their device can have Claude install stuff on their device, write code, build software, do all kinds of stuff, right. You might do stuff to your device that’ll mess things up.

49:09- Daniel:

Right. Okay. Worst case. That device has to go back in. You have to get a replacement device. you know, there’s some damage there and you might have lost some downloaded data, but the recovery method methods for that. could you build a computer virus and haven’t tried it? But that would be malicious work, right?

49:27- Daniel:

I don’t think anybody would create it by mistake. and so I think you have to kind of know what is the extent of the damage you’re capable of and be aware of it. And then at the same time, you, I think organizations should have people experiment with it. Because the real breakthrough here is not in the small increments.

49:48- Daniel:

The 99% of people who use AI use it to, Hey, clean up my email, write a blog post. Take what I’ve written and refactor it, that’s what most people use it for. A lot of people at a personal level will use it for generating ideas from the web. That’s not really interesting in the sense that it really, again, it defaults back to the, lowest common denominator from the market.

50:13- Daniel:

You’d, you’d wanna do deep research, give it very clear instructions, upload some documents and say, okay, now study this and within a certain framework, give me, you know, here are my thoughts and balance off of that or take them further. Right? But just asking a general question like what is the easier, how would one write an outbound email back to that example, because it’s so common.

50:37- Daniel:

it’s not interesting For an organization that’s very minor, incremental, improvements in productivity. The big, enormous shifts and leap forwards in productivity are things like, how do we eliminate Figma together? How do we eliminate the need to go through multiple review cycles internally? How do we eliminate the need to wait a week until we go to that next meeting where we will, everybody reports back and says, yes, this is ready for shipping.

51:08- Daniel:

And in fact, it was ready on Monday, but we have to wait till Friday meeting to hear that from you because that’s how we work. The big leap forward is you don’t have to wait. It knows, the system knows that you told somebody else in the organization, it’s ready, and I don’t have to wait till I meet with you on Friday for you to tell me.

51:26- Daniel:

I will just know right away because this AI knows, therefore I know and those things. Can only change because people experiment, because people get creative, because people dare and they dare to change something and they dare to do things, not the way we’ve always done them, and they dare to propose something

51:46- Terry:

you spoke before about the, putting an integration in place, which is effectively, you know, the next level of automation for the messaging in outreach for the sales team. For me, that’s almost the, you know, the framework you spoke about then in terms of, when is it worth automating versus, you know, when does it still need human judgment?

52:06- Terry:

It’s like when you’ve seen consistent output that you haven’t need to alter, you said for a month. do you have a particular framework in mind that you use to judge whether something is worthwhile automating or whether it still needs human judgment?

52:21- Daniel:

Yeah, I think the automation falls into a few categories and there’s no difference here between this and classic automation. The question is how much money is it costing the business? How much will it, how much will it cost me to deliver this? What’s missing? What’s different here? What’s missing here is that the cost to deliver it is a lot lower potentially than it has ever been.

52:47- Daniel:

’cause we don’t have to go and buy a product for it. We don’t have to get external people. And, we have internal resources that can do it. The challenge that we see today is it’s not clear which, internal resources are or who the we don’t have dedicated, ops teams that know how to do this well in most organizations.

53:06- Daniel:

So I think next year we’re gonna stop seeing ops teams or IT teams dedicated to internal. AI solution building, and they will, and we see some of those today, but there are very few and far between. We’re gonna see this evolve into a real team that’s standard in most organizations, even small ones. And all they will do is look around and somebody will say, Hey, can I have this workflow I wanna build.

53:33- Daniel:

Can you build it for me? And it’ll be as almost as easy as building a BI report. The, the way the bi, the op seems, you fill out the form. The next week you get a BI report saying, here, here’s what you asked for, right? it’ll be the same. They’ll go through some kind of review process. They’ll build a prototype.

53:50- Daniel:

They’ll say, Hey, does this work for you? Say, this is great. It’s a little more than that. Somebody has to maintain these things that, you know, somebody knows, you know how to troubleshoot it when it breaks all these stuff. That’s why we need a dedicated team that has the proper access. They can, they can be accountable for these things and maintaining them and keeps track of all of these different workflow AI workflows, but this, process optimization industry that when the past cost millions and we had to bring people in from the outside, it’s gonna come in house.

54:20- Daniel:

It’s gonna be very cheap, and we’re gonna see a lot more of it happen internally. You’ll just have to make a case for it and say, and it’ll be prioritized based on ROI, as simple as that, right? And some stuff we’re gonna be so simple and so easy, it’ll just be a, it’ll, somebody could just do it in five minutes, because all it takes is just, I need a prompt.

54:41- Daniel:

Can you just make that happen? Just connect these two things away with the prompt, right?

54:45- Terry:

Yeah, yeah. look, there’s a lot of conversation around governance. you’ve given some examples of practical examples, at a lower level where that you’ve put some things in place. but what does good AI governance actually look like inside an organization from an organizational level?

55:02- Daniel:

I don’t know if I’m the best person to answer that question. and I don’t think anybody knows. I don’t think there’s a good template or playbook for that yet. I think people know, but they don’t have, there’s nothing has really been fully baked probably because this thing is changing so quickly and, and probably because like I said, we want to let people dare, right?

55:26- Daniel:

And so that’s what a potential lies. And so how do you balance those? So governance, I can’t speak so much. I’m not an expert in governance and bureaucracy and legal and, you know, management of these, that type. What I can say is that having somebody in the organization that understands these systems and can advocate for them and understands their potential and their risk and understands how they work, is really, really important.

55:56- Daniel:

Because if you just ask your average CIO or IT team, they’ll have so many questions. There’ll be, or even CTO would overthink it, right? They’ll have so many questions that it’ll become analysis paralysis. They can’t move forward. But if you have an evangelist, the chief AI officer type, it’s accountable. Understands it and says, look, this is worthwhile pushing forward. It’s worthwhile taking this, these type of risks. I think that will help a lot. And then the governments, the governance question is a function of that. and they would know how to, do that, to manage that in the organization and work with the various, stakeholders to really get set it up right for your organization.

56:43- Terry:

Excellent. Yeah, I agree. I think the big point is having the expertise within the business to understand both the opportunities and the potential pitfalls. You shared a story last time we were talking about somebody blindly trusting AI to fill out a regulatory form, that approval happened, and there was some to and fro, but it raises a bigger question for me.

57:03- Terry:

What happens when AI gets it wrong and nobody’s watching? what’s your take on where oversight needs to sit in?

57:09- Daniel:

Well, the question of course is who set it up. I think it’s very similar to self-driving cars. What happens when a car drives by itself and, nobody was sitting in the driver’s wheel, right? and an accident. So if there’s no accident, nothing happened, right? The answer is nothing. Right? Everything is good.

57:30- Daniel:

The when it does, when something does go wrong. Who is to blame, right? who set this up? Who’s responsible? The answer is, well, if either the driver or the car owner, right? and if you, and if there’s no driver there, then it’s whoever owned that car and had it and hit the stop button or what, whatever that was.

57:50- Daniel:

So I think the same thing goes in an organization. It’s like if you’re just gonna create stuff and let them run, and we do this all the time, people, I knew somebody who worked in a, had a business and The entire model was, we will go into your business by all those servers and computers that are running, and they don’t actually do anything for your business.

58:12- Daniel:

We’ll shut them down one by one and we’ll save you millions. You can’t imagine how many servers are running in your business. And different computers that do nothing for your business. We’re like, yeah, we know, but we don’t know how to find them and we don’t wanna take the risk and we’ll do that for you.

58:28- Daniel:

Right? And they go in and they, now there’s a lot more is cloud-based. And you can do the same for cloud, right? There’s so many different servers running and it turns out you don’t need them in the first place. So they could be a lot more, efficiently run. And so the same question happens, right? Who set that up?

58:46- Daniel:

Who just left it there? Why didn’t you maintain it? Who’s responsible? And that’s why I think the concept of, an internal AIOps team, and that could be a team of one, it could be a larger team, is so important. They’ll be able to not only be accountable, they will be able to move stuff along, to learn, to set up a framework.

59:08- Daniel:

So, and set up, Perhaps layers of protection. So when we first came out with our ai, integration solution, so active track measures, employee activity and devices, right? We do this, we do this at a very, meta label. So we don’t read emails, we don’t watch cameras. We just kind of want to know where is productive work happening when, so businesses can make better decisions.

59:34- Daniel:

And we collect all that information and we provide different interfaces and, user interfaces and reporting tools. But we also have an API and we were able to make that AI ready and we have a whole solution for businesses that want to connect that to their AI systems so they can analyze it in multiple new ways. And one of the first things that I knew we would need would be, Hey, we need a layer between us and their ai. That is able to take all the business logic and all of the security and privacy and compliance requirements, all of these things and put that in and someplace to put all of that in a place that can be looked at, a human would be able to log in and say, yes, I reviewed all of these and I can add to this when as we learn lessons or I can release a bit as it becomes less relevant.

01:00:28- Daniel:

Or hold on a second. In Europe we need to do a little more this and that. If somebody is, accessing the data from Europe, make sure that Right. And so just being able to add that in. And I think whenever you have an internal team, they’re able to either informally or very formally, take that layer and bake it in, either bake it into whatever they do or.

01:00:51- Daniel:

Literally create a, layer where they take those requirements and write it down into their workflow. Right? And I think that’s why it’s so crucial that this will start happening soon, because the question of who’s responsible is gonna start showing up more and more,

01:01:06- Intro:

And so that, to that point, there always needs to be, from what you’re saying there, there can never really be a time when nobody’s watching, whether it’s directly or through policy and, process. There has to be someone involved at some point in time. But, you know, you said something very powerful at the beginning of this conversation and I didn’t see, I actually, made a note.

01:01:27- 

I didn’t skip over it. You said that AI is starting to shape new religions and political systems, even how we’re define truth. Where do you see this going in the next three to five years?

01:01:38- Daniel:

That is a chapter I just started writing. So, I think in the next three to five years. That’s really interesting. I think one thing that I’ll give here is that in a, fractured world, in a world where we have a infinite personal realities, right, the way I see the world and information I get about the world is different than yours. People are very uncomfortable with that. This kind of uncertainty, uncertainty and truth, that is a, an uncertainty in who and what to trust is not a place society likes to have or tolerates for a long time. In other words.

01:02:16- Terry:

Sorry, but, that’s not dissimilar to before AI though, is it? I mean, depending on which news channel you watch, you, we see different perspectives of the world. My personality and personal experiences over my life will give me a different viewpoint. We can both be looking at exactly the same. Or, event that may have happened on a world stage, and we will see it and experience it in different How is ai.

01:02:39- Daniel:

of reality has already happened and we’re seeing a lot, we’re seeing the consequences of that starting to already emerging. What AI adds to that is a much deeper lack of trust in any authority. Meaning if you can’t believe a video, an image, a sound recalling, you can’t tell truth from fiction anymore.

01:03:00- Daniel:

And. You can be validated pretty much of anything you want to be validated with, right? You can find proof for anything because you simply tweak the prompt to get it. The lack, in, the insecurity and the lack of trust in authority is getting worse.

01:03:19- Terry:

And so it’s like the, search engine algorithms on steroids. You know, you search for something and it gives you more and more of what you’re looking for, which is validation and of your particular circum or, position. Then with ai, it’s doing that to an even greater level because it’s a two-way conversation.

01:03:38- Daniel:

And so when a situation, when it’s when reality is uncertain, when you have lack of trust in authority and organizations and there’s a lot of data to show. That lack of trust in political parties in the court system and doctors in education. There’s so much showing it. You know, students today, they trust AI as better teachers than they trust our teachers. and people will trust AI more than they trust our doctors. This is incredible data that’s coming up and it’s changing very fast in the last couple of years. And so society cannot function like this for a long period of time. It does not like it. So you’re seeing a mo a few models emerging and some closer than others.

01:04:25- Daniel:

You can see the Chinese and Russian models, two different models, Chinese authoritative model where it’s, okay, we will create one reality that everyone must adhere to and we will control your thoughts almost to that level. The Russian model, which is, Hey, let’s embrace this and take advantage of this and make everything seem fake to the level of nihilism of, it doesn’t even matter anymore.

01:04:52- Daniel:

I just give up. I’ll just let the government do whatever it does, and that’s it. And I just resign from reality altogether. or in the western world, we have more free markets and, free choice and things, so we will have people choosing which solution they want. Some will, choose religious authority as an ultimate source of truth and reliability and confidence.

01:05:15- Daniel:

Some will choose, charismatic political leaders, maybe an emerging ideology or two, you know, some of those already shaping up, or evolving. And, some will turn to AI as an authoritative truth because it’s data-driven. It’s not, human. it’s all these things that in some ways we dreamed of.

01:05:38- Daniel:

And in some ways we treat AI as a deity, you know, so not just a sort of a source of authority. So I’m not saying AI is gonna be the new God, but, there is a phenomenon of data and there is a phenomena where people are slowly delegating more to AI and accepting its, authority over them, and abdicating a lot of their decision making to AI subconsciously at this point.

01:06:03- Terry:

you know, I, it is so perplexing to me in, I agree with what you’re saying in terms of people, I won’t say that it’s a faith or a diary, but they believe, they believe what’s happening, and what AI is confirming. And yet then you see there, there seems to be this naivety. Do they not think that, do you, I think society is forgetting the fact that these large, you know, these LLMs.

01:06:32- Terry:

These organizations that are building these huge engines, they have a perspective. They have, whether it’s a commercial reality, there is an objective that they have. And so look at yesterday’s. I think for me, that really crystallized yesterday when there was an announcement that Sam Altman from OpenAI and, and Mark Zuckerberg from Meta actually stated their objectives of their ai, platforms and development.

01:07:02- Terry:

And Sam Altman came out and said, you know, the whole objective, of AI is to replace humans. And,Zuckerberg came out and said the whole, objective of matters AI is to, amplify humans. and just that thought process, that’s the underlying objective of these two platforms is so different that anyone that is engaging with one or the other is going to have such a different reality that they’re living in that.

01:07:34- Terry:

Where does the balance come from? How does someone, you know, if they don’t have that level of awareness of what’s driving this from a top level, they don’t have the exposure across different platforms to have the different perspectives. What is the outcome? How do you say society doesn’t like it when won’t live there, but how, what’s the next step?

01:07:54- Terry:

where’s the change

01:07:55- Terry:

coming? What’s

01:07:55- Terry:

changeable ?

01:07:55- Daniel:

the change is never comfortable. So when you look at previous disruptions of this type, they have led to revolutions and not necessarily a violent revolution, but they’ve led to revolutions where a whole shift in the balance of power and in social structures, religious changes. and you know, the classic example is the, pro Protestant revolution, right?

01:08:22- Daniel:

the Reformation. It was a revolution. monks lost their place in society and their power. the Catholic church lost a lot of territory and, coverage, so, it’s very hard to say exactly what happened. But what we do know is that inevitably some change must happen because we’re already seeing it take place.

01:08:45- Daniel:

We’re seeing a whole new class, and, more than one class emerged the whole gig economy class. Didn’t exist before. we’re seeing a class of the tech billionaires or the tech like you said, the Zuckerbergs and Musk of the world that have not only now technological power, now they have a lot of political power.

01:09:05- Daniel:

And what they decide doesn’t just control what our phone can do. It controls our realities, right? And they now have political interests above and beyond just, hey, you know, change some regulations to benefit my company. That they’re intertwined with, media and, trade and many, many different social, policies that are directly tied to the kind of work that they do.

01:09:32- Daniel:

If robots are gonna replace humans as a Tesla. that’s what they want. Or for, you know, they, this is intertwined with, policies, right? And Uber don’t want to pay, they don’t want their gig workers to be employees because they, see themselves as a, network platform, not as, a software platform that just coordinates and moves information around.

01:09:56- Daniel:

They don’t see themselves as a cab company at all. Right? And all these things are intertwined now. And so that is a class in and of itself. And of course there’s theinfluencer class. And those include religious influencers, political influencers, and people that are, they don’t come from the traditional background of.

01:10:15- Daniel:

Authority through education and through the origin. the typical, process of they went to university or they learn, they have, religious influence because they have, they’re priests. They have religious influence because they’re charismatic and they can gain a much bigger following and have a much bigger influence and make a lot more money than a cardinal, right?

01:10:38- Daniel:

So, or any mega church for that effect. So these are shifts that are happening today already in the structures of power. And we’re seeing some of those things play out in politics today. A lot of the policies that are being debated, a lot of the, Fast, rapid shifts back and forth in the reactionary approach today to politics is a direct outcome from these changes.

01:11:07- Daniel:

It’s not just a, something that happens, right? We can, all of these things are an outcome of these fractured realities. Immigration policy, medical policy that we see today are outcomes that are driven by information, in social media and social networks. And now AI is turbocharging that.

01:11:31- Daniel:

And so when you see people relying on a trusting AI more than they trust doctors, well, when you have a, people in the government that are not doctors making. Very concrete statement, opinions and policies based off of that are aligned with those, with what people hear in social media or AI is telling them, well, they are much more okay with it.

01:12:00- Daniel:

Right? And these, I would argue, these policies are designed to, to mimic, they’re not random, they’re not people’s whims, they are not people’s ideologies. They are directly mirroring different, trends on social media, which are directly feeding into ai, training. So there’s a loop going on here between all these things and some are being manipulated, but mostly it’s, like you said, it’s the biases from within the system.

01:12:34- Daniel:

For good or bad. there are always biases, no matter there’s no system is, can be completely neutral because being neutral is biased in its own way, right? those studies that show that angry content gets I think 10 times more circulation and more reaction online than content that bring, that evokes any other emotion, right?

01:12:58- Daniel:

Angry content is associated more with right-wing content because that’s just how things work, right? And so by definition, these platforms being neutral, promote more right-wing content. And this is not a political statement or anything like that. It’s just an observation of how mechanics of things work. And so the question of biases in the system will always be there, right? And so some will argue, well, to fix that we need to create biases and some will simply be cynical and take advantage of biases. and you can see the model of, say Turkey and, you know, strong men approach Let’s bend reality around us and bend policies around us to meet our specific needs and take an opportunistic advantage of whatever trend happens here.

01:13:52- Daniel:

Sort of, you know, south America style with politics, right?

01:13:55- Terry:

I’ve just listened to everything you’ve said and you prefaced at the beginning, and you’ve said it a couple of times, that you don’t necessarily think that, you know, there’s a, that we’re approaching a new reformation, but I’m getting the sense that I think you do, and of all the people I’ve spoken to, I get a sense that you’ve, to a much deeper level, given it much more thought certainly than anybody I’ve spoken to before.

01:14:17- Terry:

Now, the reformation, not necessarily of religion, but of power and truth, which is exactly what you’ve just described, we’re currently going through now. what do you think that might look like in our lifetime? what do you think the future that this represents or that pre will present for us?

01:14:33- Daniel:

I think we’re gonna see a lot more uncertainty before we start seeing certainty. And I think we’re in a very unique and unprecedented period in history because never before have we seen two and possibly three, major communication, I would say slash technological slash communication.

01:14:53- Daniel:

disruptions of this magnitude happen so close together that the repercussions of one haven’t been completed before the next one started. And so, you know, the digital revolution. We’ve all seen a known with the, social media communication and all that. And even that hasn’t completed yet as we’ve gone from COVID and everything that’s still evolving.

01:15:15- Daniel:

And we’ve had, you know, the great deep platforming, for example, that’s very recent, that, al many of the AltRight movements and the far right movements have moved off the mainstream platforms into their own. Platforms, and that’s gab and through social, et cetera. And what are the repercussions of that?

01:15:32- Daniel:

That even hasn’t said it yet. So this is still new. Even if you’re saying, yeah, that is, that isn’t that already the truth, that already the situation where it’s still evolving and happening, right. And society hasn’t adjusted, government systems haven’t adjusted. And if you look at what happened to the Democratic party in the US that is all, that collapse of that party today, they have no clue how to get out of that situation they’re in right now.

01:15:57- Daniel:

That is directly correlated with, them not knowing how to deal with this new types of communications. it is not their platform. It is their inability to work in real time against these new types of communications. And they just can’t handle it. They need to figure it out. and it so happened that the Republican party was able to embrace these things a lot faster.

01:16:20- Daniel:

And a lot better. So that’s still a play. Now we’re having AI starting just starting right, and it’s moving so fast. And in two, three years, we’re gonna start seeing the robotics revolution come in. And that’s gonna be mind boggling because that’s gonna affect class in a way that we haven’t seen before because everybody that has been affected by globalization and offshoring of jobs, so we’re talking about industrial work, et cetera, that now are being promised by right wing, sort of, conservative movements will bring back the glory of the old days before globalization and bring back old, industrial work to the US.

01:17:03- Daniel:

And this is echoed across the world, similar claims, right? whether that is a good idea or not, whether that will succeed or not, before we will even know. Robotics are gonna disrupt that very same population disproportionately more than anyone else. These are factory. if we bring back factory workers to the US and we have more expensive labor, well what is gonna happen?

01:17:26- Daniel:

We’re, there’s gonna be a much stronger incentive to bring, to build them off of robotics than before. So it’s gonna accelerate that movement and the impact on society and the impact, and the disillusionment once again is gonna be even stronger. And so history tells us that in such periods of change, with one disruption, you go through different cycles of, could be violent, could be nonviolent, could very, let’s say, uncomfortable shifts in power and evolution of political and religious thinking. Now we’re talking about three of them. So I do not dare to say exactly what happened when we see new religions come out. I think it’s way too early to know exactly what happened. I think religion will adapt and change and we might see some emerge. They may or may not succeed, but I wouldn’t venture to guess what is the ultimate outcome.

01:18:23- Daniel:

But I will say we’re gonna see a lot more discomfort, social discomfort, and it’s gonna lead to a lot more, very charismatic figures saying, I’ve got the answer. We’ve gotta put a foot down and create one social order. they’ll put an end to this discomfort. We haven’t seen as much of that from the left as we have from the right, but it will come.

01:18:45- Daniel:

It must come. And when you have two opposing such, charismatic figures duking it out in public with a big following that nowadays you can bring out into the street a lot more easily and rapidly than before. The cycles around that the ability to, mobilize is unlike anything we’ve seen in the past. we’re gonna see more of that, right, which will enhance the cycle of the demand for a one true social order, right? So, that is the worst case scenario is that that actually happens, that takes place and we end up like China, right? The, the better scenario, which I’m optimistic about and I think is more indicative of the Western approach to the world, is.

01:19:31- Daniel:

The free market will allow for multiple different ways to handle this. The market will handle it, and you’ll see religious ais pop up and you’ll see dictatorial ais pop up and you’ll see all kinds of, political systems pop up and your religions and different things. So people will find comfort in different ways and in the end, it’ll balance itself out by itself, hopefully without too much cost and without too much heartache along the way.

01:19:59- Daniel:

And if it does pan out, we will be in a much better place compared to any previous generation. We will, at the end of it, we will be very, very happy. but getting there is not easy.

01:20:11- Terry:

Wow. Okay. very, very interesting landscape that you’ve painted and, look great, great food for thought. I think it’s, insightful. I’ll need some time to think about that, actually. I’ll need some time to think about that one.

01:20:23- Daniel:

But that is a great compliment and I appreciate it.

01:20:26- Terry:

Hmm, hmm. I’m gonna ask you a couple of rapid fire questions now. getting close to a wrap up.

01:20:32- Daniel:

what skill will be the most valuable for marketers and leaders over the next five years?

01:20:37- Daniel:

wow. data, understanding data, how to use it, how to work with AI on data. does not do data well yet. That’s gonna come, but that’s gonna be a lot of trial and error associated with sort of self-serve bi work, with AI as an intermediary. As it is today, we see examples of you feed AI a spreadsheet and ask it to analyze it.

01:21:02- Daniel:

It’ll, you think it knows how to do it. It does it, it knows how to write code or it knows how to guess code that can analyze data, but it has no clue what it’s doing. It doesn’t understand data, it understands statistical probability of what should be the next word in a sentence and the next word after that, and the next word after that.

01:21:20- Daniel:

So, as we have more and more data as we need, the expectation is anybody can generate data, or reporting with AI quickly, without having to go through longer loops and go to other people in the organization. Well, you have the stuff connected, you have ai, just figure it out, right? there is a lot more need for everybody to understand data and to understand how to spot.

01:21:43- Terry:

Hey, no, no, no, no. You just eliminated half my spreadsheet. You didn’t tell me. And that does happen now pretty much consistently. So, that is the kind of stuff I think really everybody should know. Yeah. And data is truth. whereas what comes out of ai, everything needs to be validated. Whereas if you’ve got the base data and know how to use that to form your base opinions and base, expectations and to make decisions from then, it’s almost a validator for anything that might come in terms of AI interpretation from there.

01:22:14- Terry:

I think that’s a, that’s an excellent point. what’s one practical piece of advice you’d give a team? trying to integrate AI without losing trust?

01:22:23- Daniel:

practical. Wow. try not to mess it up too much. So don’t give a good excuse for somebody to shut you down. I think being transparent about how you do things, don’t try to hide it. Like, some people will use AI and sort of hide the fact because they don’t want people, they wanna take credit for the work.

01:22:41- Daniel:

I think the true credit should come from, look how I integrated AI and look how I did this and how I leveraged AI so smartly here. Bring as much transparency to what you’re doing. Show other people. Let them see that they always can see what you’re doing and they don’t have anything to fear. I think the biggest fear comes from, what the heck is Daniel doing over there with ai?

01:23:04- Daniel:

I need to put some guardrails on him because I don’t understand.

01:23:07- Terry:

Mm-hmm.

01:23:08- Daniel:

The more you open up and the more you show, the less resistance you’re gonna have and the more buy-in because the potential is so big.

01:23:16- Terry:

Excellent. Excellent. And so then please finish this sentence. The companies that survive the AI wave will be the ones who

01:23:23- Daniel:

It dramatically changed our business to improve their profit margins compared to the competitors.

01:23:31- Terry:

got it. Excellent.

01:23:32- Daniel:

Is there a particular preference that you have in terms of people getting in touch with you?

01:23:36- Daniel:

I think I love LinkedIn. I’m always on LinkedIn And I think that’s a wonderful way to keep in touch over time. So yeah, LinkedIn is great.

01:23:45- Terry:

Yeah, we’ll have that link down there

01:23:46- Daniel:

Well, Daniel, I’ve gotta say thank you very much. It’s been a long conversation this morning, but one that certainly, has been absolutely insightful. I’ve enjoyed it very much so appreciate you coming on the chat.

01:23:55- Daniel:

Thank you so much for having me. This was wonderful and I appreciate it.

01:23:59- Terry:

Excellent. Have a great day. Bye

01:24:00- Daniel:

Bye.